How I Automated Pinterest Photo Prep With Make

 I use Pinterest quite often, and I take a lot of photos of my balcony garden. Sometimes I go through them and find a few I would like to post. Then I get to the part where I have to write a title and description for every photo, choose a board, and add a link to my blog. For one photo, that is fine. For a whole folder, it starts to feel like a job I will do another day.

I used AI to help draft this article. This post also contains a Make affiliate link. I may earn a commission if you sign up through it and buy an eligible subscription.

I wanted to see if I could automate some of that work. So I started with one photo of a pepper plant in my greenhouse and tried to turn it into a Pinterest draft.



Pinterest Can Already Read a Spreadsheet

The first thing I found was that Pinterest lets business accounts upload multiple pins from a CSV file. Each row holds the information for one pin. I could put the title, description, board, and blog link in the file, and I could add a date later if I wanted to schedule it. Pinterest’s bulk upload guide

There was one detail I had to figure out first. Pinterest needs a public link to the actual photo. The image on my phone had a filename, but Pinterest had no way to get to it.

I made a folder in Cloudinary called pinterest and uploaded the pepper photo. Cloudinary gave me a public URL I could open directly in my browser. That solved the image part. I still needed something to write the draft and put all the fields in the right places.

What I Used Make For

Make lets you connect apps and decide what information passes from one step to the next. It calls the whole workflow a scenario and the individual steps modules. In my case, that meant one module could find the photo in Cloudinary, and the next could use the URL it found to download it. I added my Cloudinary credentials in Make’s connection settings and kept the API secret there. Make’s guide to scenarios

Once Make had the image, I passed it to its AI agent. I asked for a title and a short description based on what was visible in the photo. I used Make’s own AI provider for this test. You can also connect a supported model from OpenAI or Anthropic, depending on your plan. Make’s AI agent documentation

This is one of the titles it gave me:

Pepper plant thriving in a compact balcony greenhouse

It got the pepper plant and the greenhouse right. I would probably take out “thriving,” though. You can see red peppers in the picture, but I know how the plant was doing over the rest of the season. I’m happy to let Make give me a first draft and make that last edit myself.

My First CSV Looked Fine in Make

Make took the title and description and combined them with the photo URL, my Balcony Gardening board, and a link to my blog. The two links have different jobs. The first tells Pinterest where to get the picture. The second is where someone goes when they click the pin.

The first run looked fine until I downloaded the file. I had entered a formula in the upload step, and the exported file contained the formula as text. I changed the upload step to use the CSV output directly and ran it again.

This time, I opened the file and found five columns and one complete row for the pepper photo. That was the first point where I could sit down, read the title and description next to the other pin details, and decide what I wanted to change.

What Comes Next

So far, I have tested the workflow up to the CSV export. The next step is to try that file in Pinterest’s bulk uploader and see how the pin comes out.

If I start running this regularly, I want Make to remember which photos it has already exported. One way to do that would be to tag each photo in Cloudinary after its CSV has been saved. Then, when I add more photos to the folder, Make can look for the ones without that tag. I could also try sending approved drafts to Buffer for scheduling later. Make’s Buffer integration

For now, I have a folder full of possible posts and one draft I can actually check. If you have a folder like that too, start with a photo you know well. You’ll be able to tell pretty quickly which details the workflow gets right and which ones you’d describe differently. Try Make using my affiliate link.

Comments